Description of problem:
The ES copy test script run failed without OPS clusters, full logs: http://pastebin.test.redhat.com/361648

It passed when OPS clusters are deployed in logging project. Passed log here: http://pastebin.test.redhat.com/361714

Version-Release number of selected component (if applicable):
openshift v1.1.5-66-g9b04353
kubernetes v1.2.0-36-g4a3f9c5
etcd 2.2.5

How reproducible:
Always

Steps to Reproduce:
1. Deploy CEFK stacks without OPS cluster, after all logging pods are running, check that fluentd pod logs did not contain errors
2. Prepare frontend service in -n test
$ oc new-project test
$ oc new-app https://raw.githubusercontent.com/openshift/origin/master/examples/sample-app/application-template-stibuild.json
Wait for the 2 frontend pods running:
$ oc get po | grep frontend
frontend-1-s4sqz            1/1       Running     0          3m
frontend-1-t7kv5            1/1       Running     0          3m
3. Specify --config argument in env :
$ export KUBECONFIG=/etc/origin/master/admin.kubeconfig
4. Add the oc client user to cluster admin
$ oadm policy add-cluster-role-to-user cluster-admin xiazhao
5. Go to openshift master node, oc login to openshift with user xiazhao,  and switched to the logging project, then run the test script:
wget https://raw.githubusercontent.com/openshift/origin-aggregated-logging/master/hack/testing/test-es-copy.sh
chmod 755 test-es-copy.sh
VERBOSE=1 ./test-es-copy.sh false

Actual results:
The ES copy settings was not recovered, test script run failed without OPS clusters, full logs: http://pastebin.test.redhat.com/361648

Expected results:
The ES copy settings should recover

Additional info: